Revert "[Feat] Unquantized linear nz support (#2619)" (#2896)

### What this PR does / why we need it?
This reverts commit 7b2ecc1e9a.

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
CI passed

- vLLM version: main
- vLLM main:
64d90c3e4f

Closes: https://github.com/vllm-project/vllm-ascend/issues/2890
Closes: https://github.com/vllm-project/vllm-ascend/issues/2887
Closes: https://github.com/vllm-project/vllm-ascend/issues/2885

Signed-off-by: Yikun Jiang <yikunkero@gmail.com>

tests/ut/ops/test_linear.py

@@ -5,13 +5,11 @@ from unittest.mock import MagicMock, patch
 
 import torch
 
-from tests.ut.base import TestBase
 from vllm_ascend import ascend_config
 from vllm_ascend.distributed import parallel_state
 from vllm_ascend.ops.linear import (AscendColumnParallelLinear,
                                     AscendMergedColumnParallelLinear,
-                                    AscendRowParallelLinear,
-                                    AscendUnquantizedLinearMethod)
+                                    AscendRowParallelLinear)
 
 
 class BaseLinearTest(unittest.TestCase):
@@ -48,81 +46,6 @@ class BaseLinearTest(unittest.TestCase):
             p.stop()
 
 
-class TestAscendUnquantizedLinearMethod(TestBase):
-
-    def setUp(self):
-        self.method = AscendUnquantizedLinearMethod()
-
-    @mock.patch("torch_npu.npu_format_cast")
-    @mock.patch("torch.version")
-    def test_process_weights_after_loading_is_cann_8_3(self, mock_version,
-                                                       mock_format_cast):
-        layer = mock.MagicMock()
-
-        mock_version.cann = "8.3.RC1"
-        self.method.process_weights_after_loading(layer)
-        mock_format_cast.assert_called_once()
-
-    @mock.patch("torch.version")
-    def test_process_weights_after_loading_not_cann_8_3(self, mock_version):
-        layer = mock.MagicMock()
-
-        mock_version.cann = "8.2.RC1"
-        # Should not raise exception
-        self.method.process_weights_after_loading(layer)
-
-    @mock.patch("torch.matmul")
-    @mock.patch("torch.version")
-    def test_apply_with_bias_is_cann_8_3(self, mock_version, mock_npu_matmul):
-        layer = mock.MagicMock()
-        layer.weight = torch.randn(128, 256)
-
-        x = torch.randn(32, 128)
-        bias = torch.randn(256)
-
-        expected_y_output = torch.randn(32, 256)
-        mock_npu_matmul.return_value = expected_y_output
-
-        mock_version.cann = "8.3.RC1"
-        output = self.method.apply(layer, x, bias)
-
-        expected_y_output += bias
-        self.assertTrue(torch.equal(output, expected_y_output))
-
-    @mock.patch("torch.matmul")
-    @mock.patch("torch.version")
-    def test_apply_without_bias_is_cann_8_3(self, mock_version,
-                                            mock_npu_matmul):
-        layer = mock.MagicMock()
-        layer.weight = torch.randn(128, 256)
-
-        x = torch.randn(32, 128)
-
-        expected_y_output = torch.randn(32, 256)
-        mock_npu_matmul.return_value = expected_y_output
-
-        mock_version.cann = "8.3.RC1"
-        output = self.method.apply(layer, x)
-
-        self.assertTrue(torch.equal(output, expected_y_output))
-
-    @mock.patch("torch.nn.functional.linear")
-    @mock.patch("torch.version")
-    def test_apply_not_cann_8_3(self, mock_version, mock_npu_linear):
-        layer = mock.MagicMock()
-        layer.weight = torch.randn(128, 256)
-
-        x = torch.randn(32, 128)
-
-        expected_y_output = torch.randn(32, 256)
-        mock_npu_linear.return_value = expected_y_output
-
-        mock_version.cann = "8.2.RC1"
-        output = self.method.apply(layer, x)
-
-        self.assertTrue(torch.equal(output, expected_y_output))
-
-
 class TestAscendRowParallelLinear(BaseLinearTest):
 
     def test_mlp_optimize(self):

tests/ut/quantization/test_quant_config.py

@@ -4,10 +4,10 @@ import torch
 from vllm.attention.layer import Attention
 from vllm.model_executor.layers.fused_moe import FusedMoE
 from vllm.model_executor.layers.fused_moe.config import FusedMoEConfig
-from vllm.model_executor.layers.linear import LinearBase
+from vllm.model_executor.layers.linear import (LinearBase,
+                                               UnquantizedLinearMethod)
 
 from tests.ut.base import TestBase
-from vllm_ascend.ops.linear import AscendUnquantizedLinearMethod
 from vllm_ascend.quantization.quant_config import (AscendKVCacheMethod,
                                                    AscendQuantConfig)
 from vllm_ascend.utils import ASCEND_QUANTIZATION_METHOD
@@ -79,7 +79,7 @@ class TestAscendQuantConfig(TestBase):
                           'is_layer_skipped_ascend',
                           return_value=True):
             method = self.ascend_config.get_quant_method(linear_layer, ".attn")
-            self.assertIsInstance(method, AscendUnquantizedLinearMethod)
+            self.assertIsInstance(method, UnquantizedLinearMethod)
 
         # Test quantized layer
         with patch.object(self.ascend_config, 'is_layer_skipped_ascend', return_value=False), \